EP0798889B1

EP0798889B1 - Error control method and error control device for digital communication

Info

Publication number: EP0798889B1
Application number: EP96116817A
Authority: EP
Inventors: Shinji Uebayashi; Hui Zhao
Original assignee: Nippon Telegraph and Telephone Corp; NTT Mobile Communications Networks Inc
Current assignee: NTT Docomo Inc; Nippon Telegraph and Telephone Corp
Priority date: 1996-02-29
Filing date: 1996-10-18
Publication date: 2005-05-25
Anticipated expiration: 2016-10-18
Also published as: CA2187564C; US6134694A; CN1159109A; EP0798889A3; EP0798889A2; CA2187564A1; KR970063971A; JPH09238125A; DE69634770D1; DE69634770T2; CN1083184C; KR100227351B1

Description

BACKGROUND OF THE INVENTION Field of the Invention

The present invention relates to an error control method and an error control device for digital communication which employ retransmission control such as ARQ (Automatic Repeat Request), and in particular relates to an error control method and an error control device for digital communication which use hybrid ARQ which carries out code combining.

Background

In conventional retransmission control using a code combining technique, all received signals are used in a decoding. The technique of retransmission control is disclosed, for example, in IEEE Trans. Commun. vol.33, pp,385-393, May, 1985. There is described a code combining technique using Maximum-Likelihood decoding for combining an arbitrary number of noisy packets. Repeated packets are encoded with a code rate R. When a packet is first received it is decoded by a rate R decoder. The CRC bits are checked after coding. If errors are still detected, this packet is combined with the next received packet by using a R/2 rate decoder. If errors are still detected, a third packet is combined by using a R/3 rate decoder, and so on. Each packet is weighed by an estimate of its reliability. Retransmission control methods using code combining techniques are classified into methods which add only an error detecting code to the first signal transmitted without providing a redundancy for error correction, and methods which provide a redundancy for error correction to the first signal transmitted.

In retransmission control using the code combining technique, when signals having extremely many errors are included in the received signals, multiple retransmission is necessary because all the received signals are used in the technique. Furthermore, in the method which adds an error detecting code only to the first transmitted signal without providing a redundancy for error correction, the first signal received can be decoded only if it is received error-free. However, if there is an error in even one bit, then this signal cannot be decoded. On the other hand, in the method which provides a redundancy for error correction to the first transmitted signal, the throughput is improved in the case where the transmission channel is in poor quality. However, in the case where the transmission channel is in sufficient quality, the throughput plateaus.

If Chase decoding is used on retransmission control using a code combining technique, it is possible to correct errors using only an error detecting code. However, many combinations must be calculated in order to carry out Chase decoding, so that the number of calculations becomes enormous.

In IEEE Transactions on Communications, vol. 43, (1995) June, No. 6, New York, US, pages 2005-2009, there are described complementary punctured convolutional (CPC) codes. A starting code rate can be chosen to match the channel noise requirements, and packets that are detected in error are not discarded, but are combined with complementary transmissions provided by the transmitter to recover the transmitted message. To each information packet to be transmitted are appended parity bits for error detection and known tail bits corresponding to the memory of the encoder. Viterbi decoding using perforation pattern is applied on the received sequence. The decoder sequence, which corresponds to the information and parity check bits, is then examined by the error detection decoder. If this sequence is declared error-free, transmission of the information packet is complete. Otherwise, that sequence is saved for subsequent decoding attempts and the information packet moves to the next level. In the subsequent levels, coded bits are sent to the receiver. Viterbi decoding is first applied on the received sequence, using perforation pattern. If the decoded sequence is declared error-free, transmission of the information packet is complete. Otherwise, Viterbi decoding is applied once again, but on the combination of the sequences received up to the corresponding level, using a combined code. If decoding is successful, transmission of the information packet is complete. Otherwise, the sequences received so far are saved and the information packet moves to the next level. At last, if decoding is not successful, all received sequences available at the receiver are decoded.

SUMMARY OF THE INVENTION

It is accordingly a primary object of the present invention to provide an error control method and an error control device for digital communication, that are able to decode with high efficiency, and are able to obtain a maximum throughput even if the transmission channel has a low quality

The object is solved by the features of the independent claims. The dependent claims are directed to preferred embodiments of the invention.

Therefore, in an error control method and an error control device for digital communication, it is possible to decode with high efficiency even if the received signals contain a signal which has extremely many errors because the decoding procedure is carried out using only a portion of the received signals. Furthermore, because the signal to be transmitted has the redundancy of only an error detection code, it is possible to obtain the maximum throughput when the transmission channel is of high quality. Furthermore, due to Chase decoding, it is possible to correct errors and maintain high transmission effects even if the quality of the transmission channel deteriorates. In addition, it is possible to reduce the amount of signal operations on the receiving side by controlling the decoding order in Chase decoding.

BRIEF DESCRIPTION OF THE DRAWINGS

Further objects and advantages of the present invention will be apparent from the following description, reference being made to the accompanying drawings wherein preferred embodiments of the present invention are clearly shown.

In the drawings:

Fig. 1 is a block diagram showing a structural example of a receiving apparatus according to an embodiment of the present invention;

Fig. 2 is a chart diagram explaining the operation of a transmission apparatus and a receiving apparatus; and

Fig. 3 is a flowchart showing the Chase decoding method of the present invention.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

An explanation will now be made of preferred embodiments of the present invention with reference to figures.

Fig. 1 is a block diagram showing a structural example of a receiving apparatus according to an embodiment of the present invention.

Fig. 2 is a chart diagram explaining the operation of a transmission apparatus and a receiving apparatus. Transmission apparatus 20 receives transmission information from transmission terminal 10, and then adds an error detecting code (CRC code) to the transmission information. A signal series added CRC code is expressed as I. Next, transmission apparatus 20 carries out a convolutional encoding to the signal series including CRC bits at a rate of 1/2. The generator polynomials are expressed as (G1, G2). Furthermore, a signal encoded with G1 is expressed as P1, and a signal encoded with G2 is expressed as P2. Transmission apparatus 20 transmits the transmission signal P1 first.

Receiving apparatus 30 receives signal P1 in which errors are overlapped thereto on the transmission channel. The overlapping error pattern is expressed as E1, and the receiving signal is expressed as R1. Therefore, receiving signal R1 becomes a value (R1=P1+E1) which is sum of transmission signal P1 and a moduls-2 of error pattern E1. Receiver 31 of receiving apparatus 30 receives receiving signal R1, and then converts receiving signal R1 to a base band signal, and outputs the base band signal to receiving signal buffer 32. Control portion 35 detects receiving signal R1, and then instructs an execution of a CRC decoding procedure to error correction decoding portion 33. Error correction decoding portion 33 receives the base band signal from receiving buffer 32, and carries out CRC decoding in CRC decoding portion 331 (Fig. 2 (S1)). If CRC decoding portion 331 can decode the base band signal, receiving apparatus 30 outputs a decoded signal to receiving terminal 40 through interface portion 36, and transmits a receipt acknowledgment signal ACK to transmission apparatus 20 through transmission portion 37, completing the receiving procedure. On the other hand, if CRC decoding portion 331 cannot decode the base band signal because of errors, receiving apparatus 30 carries out Chase decoding by Chase decoding portion 332.

Fig. 3 is a flowchart showing the Chase decoding method of the present invention. Chase decoding portion 332 checks whether errors exist in the receiving signal or not (S20). When errors exist in the receiving signal, Chase decoding portion 332 searches the reliability of each bit of the receiving signal, and selects t bits which have a low degree of reliability. It is assumed that errors were generated in all or a portion of these t bits, and that errors where not generated in the remainder of the bits. There are 2^t-1 (^ expresses the power of the number) error patterns in t bits. Next, Chase decoding portion 332 calculates the sum of the reliabilities of the bits in which error has occurred in the error patterns, and carries out CRC decoding in the order of the patterns in which the sum of the reliabilities is a small value (S22).

When CRC decoding is successful (S24), error correction decoding portion 33 completes the Chase decoding procedure. Next, when Chase decoding is completed, error correction decoding portion 33 transmits receipt acknowledgment signal ACK to transmission apparatus 20, and completes the receiving procedure (S26). On the other hand, when CRC decoding fails, error correction decoding portion 33 judges whether or not there is a combination which has not been tested (S28). Furthermore, when CRC decoding for all combinations fail, and all attempts are completed (S28), error correction decoding portion 33 transmits retransmission request signal NAK to transmission apparatus 20 (S30). On the other hand, when all attempts have not been completed, error correction decoding portion 33 generates the pattern of the next attempt by reversing the appropriate bits (S32), and returning to step S24.

When transmission apparatus 20 receives retransmission request signal NAK, it transmits transmission signal P2. Afterward, transmission apparatus 20 alternately transmits transmission signals P1 and P2 whenever it receives retransmission request signal NAK. When receiving apparatus 30 receives receiving signal R2 obtained by adding transmission signal P2 and error pattern E2 (R2=P2+E2), CRC decoding portion 331 begins by carrying out CRC decoding (S3). At this time, if CRC decoding portion 331 cannot decode, then receiving apparatus 30 carries out Chase decoding by using Chase decoding portion 332 (S4). Furthermore, if Chase decoding by Chase decoding portion 332 also fails, receiving apparatus 30 carries out Viterbi decoding using R1 and R2 in Viterbi decoding portion 333 (S5). If Viterbi decoding by Viterbi decoding portion 333 fails, then receiving apparatus 30 transmits retransmission request signal NAK.

Next, receiving apparatus 30 receives receiving signal R3 which is obtained by adding transmission signal P1 and error pattern E3 (R3=P1+E3), and carries out CRC decoding (S6). If CRC decoding fails, then receiving apparatus 30 carries out Chase decoding (S7). If Chase decoding also fails, receiving apparatus 30 carries out Viterbi decoding using R1, R2 and R3 (S8). If Viterbi decoding fails, then receiving apparatus 30 carries out Viterbi decoding using R2 and R3 (S9). If this decoding also fails, receiving apparatus 30 carries out diversity decoding in diversity decoding portion 334 (S10) using R1 and R3.

Chase decoding is also used during diversity decoding. For example, when R 1 and R3 are received, diversity decoding portion 334 obtains a signal RD which is decoded by diversity decoding on the basis of a selection diversity or maximum rate synthetic diversity for each bit of R1 and R3. Next, diversity decoding portion 334 carries out Chase decoding on signal RD.

If all decoding fails, then receiving apparatus 30 transmits retransmission request signal NAK. Then, receiving apparatus 30 carries out CRC decoding and Chase decoding whenever it receives a signal, and then carries out Viterbi decoding and diversity decoding when CRC and Chase decoding fails by combining the received signals in a predetermined order. Signal selecting portion 34 decides the combination of the received signals to be decoded.

Next, an explanation will be made of the combination order of the received signals when carrying out Viterbi decoding and diversity decoding. In this example, it is assumed that there are N number of received signal series R1, R2, ...., Rn (n=2k); that the received signal series R1, R3, ...., R2k-1 is the received signal when P1 is transmitted; and that received signal series R2, R4, ...., R2k is the received signal when P2 is transmitted. In this example, it is also assumed that the reliability of received signal Rn is rn. The reliability rn can be expressed, for example, as the sum of the receiving levels of each bit in received signal Rn.

First, Viterbi decoding is carried out using all received signals. Next, Viterbi decoding is carried out using (N-1) received signals by excluding one received signal from the N received signals. There are N combinations for (N-1) number of signals. Thus, signal selecting portion 34 decides the selection order by using reliability rn. Signal selecting portion 34 selects (N-1) signals by sequentially excluding received signals Rn starting with the signal which has the smallest reliability rn. However, it is also possible for signal selecting portion 34 to select (N-1) signals by sequentially excluding signals beginning with the oldest received signals.

When Viterbi decoding using all combinations of (N-1) signals fails, Viterbi decoding is carried out using (N-2) signals. First, (N-2) signals are selected by excluding the two signals which have the smallest reliability rn. Then, rn+rn' (1=<n, n'=<N, n<n') is calculated, and (N-2) signals are selected by excluding signals Rn and Rn' in the order of smallest rn+rn', and Viterbi decoding is carried out using the selected signals.

Additionally, it is noted here that the reliability may also be calculated for all the signal combinations. For example, the reliability of received signal Ra x Rb x Rc x....x Rx is ra+rb+rc+....+rx. In this case, the reliabilities for all signal combinations are calculated, and then Viterbi decoding is carried out by sequentially selecting in the order of the combinations having the highest degree of

Then, diversity decoding is carried out on the received signal which corresponds to P1 in the selected combination. Next, diversity decoding is carried out on the received signal which corresponds to P2. Finally, Viterbi encoding is carried out by using the two signals which are obtained by each diversity encoding.

(Other Embodiment)

In this example, the number of received signals which correspond to P1 is n1, and the number of received signals which correspond to P2 is n2. If n1 is larger than n2 (n1>n2), diversity decoding is carried out for (n1-n2+1) signals having low reliability from among the received signals which correspond to P1. When this signal and the received signals which corresponds to the remaining P1 are combined, signals which corresponds to n2 number of P1 are obtained. As a result, Viterbi encoding can be carried out using a total 2 x n2 number of signals.

Claims

An error control method for digital communication which corrects errors which occur in the transmission of signals (R1-R4), comprising
a receiving step for receiving said signals (R1-R4);
a selecting step for selecting a portion of said signals (R1-R4) which are received in said receiving step,
a first decoding step for decoding said received signals (R1-R4);
characterised by
a second decoding step for carrying out Chase decoding on the signals (R1-R4) whose decoding by said first decoding step failed;
a retransmission request step for transmitting a retransmission request of said signals (R1-R4) in the case that decoding by said second decoding step fails;
a repeating step for repeating procedures of said receiving step, and said first and second decoding steps in the case that decoding by said second decoding step fails;
said selecting step including a first selecting step for selecting all signals (R1-R4) which are received in said receiving step, a calculating step for calculating the reliabilities of each signal, and a second selecting step for successively selecting a signal combination which has higher reliability from said signals which are received in said receiving step by sequentially excluding the received signals with the lowest reliability, and
said method further comprising a third decoding step for, when said repeating step fails, decoding using said portion of said signals selected by said selecting step.
An error control method according to claim 1, wherein said third decoding step carries out Viterbi decoding.
An error control method according to claim 1, wherein said third decoding step carries out diversity decoding.
An error control method according to claim 2, comprising a diversity decoding step, which is performed when said third decoding step fails.
An error control method according to any one of claims 1 to 4, comprising
a bit position selecting step for selecting a bit position from the received signal which has a low degree of reliability, by means of a predetermined bit number from the received signal; and
a generating step for generating error patterns by assuming that said errors exist at said bit position;
wherein said second decoding step carries out Chase decoding for said signal on the basis of each error pattern.
An error control method according to claim 5, wherein said second decoding step calculates the reliability of each error pattern by using the reliability of each bit of the received signal, and successively carries out Chase decoding on the received signal in the order of the error patterns which have the highest reliability.
An error control device for correcting errors which occur in the transmission of signals (R1-R4), comprising
a receiving means (30) for receiving said signals (R1-R4);
a selecting means (34) for selecting a portion of said signals (R1-R4) which are received by said receiving means,
a first decoding means (331) for decoding said received signals (R1-R4);
characterised by
a second decoding means (332) for carrying out Chase decoding on the signals (R1-R4) whose decoding by said first decoding means (331) failed;
a retransmission request means (33) for transmitting a retransmission request of said signals (R1-R4) in the case that decoding by said second decoding means (332) fails;
a repeating means for repeating procedures of said receiving means, and said first and second decoding means in the case that decoding by said second decoding means (332) fails;
wherein said selecting means is adapted to select all signals (R1-R4) which are received by said receiving means, to calculate the reliabilities of each signal, and to successively select a signal combination which has higher reliability from said signals which are received by said receiving means by sequentially excluding the received signals with the lowest reliability, and
wherein said device comprises a third decoding means for, when said repeating procedures carried out by said repeating means fail decoding using said portion of signals selected by said selecting means.
An error control device according to claim 7, wherein said third decoding means (333) is adapted to carry out Viterbi decoding. .
An error control device according to claim 7, wherein said third decoding means (333) is adapted to carry out diversitiy decoding.
An error control device according to claim 8, comprising a diversity decoding means (334), which is adapted to perform diversity decoding when said decoding by said third decoding means (333) fails.
An error control device according to any one of claims 7 to 10, comprising
a bit position selecting means for selecting a bit position from the received signal which has a low degree of reliability, by means of a predetermined bit number from the received signal; and
a generating means for generating error patterns by assuming that said errors exist at said bit position;
wherein said second decoding means (332) is adapted to carry out Chase decoding for said signal on the basis of each error pattern.
An error control device according to claim 11, wherein said second decoding means (332) is adapted to calculate the reliability of each error pattern by using the reliability of each bit of the received signal, and to successively carry out Chase decoding on the received signal in the order of the error patterns which have the highest reliability.